Discussion
We have learned significant lessons from this unique complication. Iatrogenic VSD in the setting of systemic RVAD causes significant hypoxia due to the significant inflow suction by the VAD from the pulmonic ventricle, resulting in a right-to-left shunt. The iatrogenic VSD was effectively treated with VSD patch closure, a similar technique that is employed for a post-infarction VSD. The second significant challenge was persistent right-to-left shunt and desaturation despite having adequate VSD patch and a small patch leak, highlighting the clinical significance of even small shunts after VAD implantation. This is likely due to the combination of high negative pressure from the inflow cannula in the morphologic RV and dysfunctional and distended subpulmonary left ventricle. The mechanism of the remarkable effect of iNO in our case is somewhat unclear, but perhaps explained by reduction of the afterload and subsequent improved ventricular ejection and decompression of the subpulmonary LV.
Determining the favourable positioning of the inflow cannula in ccTGA is a technical challenge. Using routine TEE offers guidance for cannula insertion in most VAD cases, but was not sufficient in preventing iatrogenic injury in this case. The authors would recommend more detailed pre-planning using cross-sectional imaging, such as computed tomography or magnetic resonance imaging, and the addition of epicardial studies for seating the VAD inflow. Although extremely rare, iatrogenic VSDs have been well reported in adults.2,3 Barioliet al describes acute VSD creation during septal myectomy for hypertrophic obstructive cardiomyopathy that was initially repaired intraoperatively with a large bovine pericardial patch, with delayed percutaneous closure of residual VSD using the Amplatzter multi-fenestrated septal occluder.4 Successful transcatheter iatrogenic VSD closure after aortic valve replacement has also been well described.5 In our case, the acute desaturation from a large iatrogenic VSD was managed intraoperatively with patch repair, RVAD revision, and the addition of supplementary iNO. The possibility of percutaneous intervention for the residual shunt was discussed. However, considering the technical difficulty of device closure near a VAD inflow and the relative improvement of the patient’s clinical status with iNO, further procedural intervention was deferred. The presence of the VSD patch and residual shunt did necessitate an earlier listing for heart transplantation postoperatively compared to patients with uncomplicated VAD placement. This case highlights a unique iatrogenic complication during VAD implantation in a patient with complex congenital heart disease, and the management steps that were taken to troubleshoot hemodynamic instability.